Joe Hockey will be desperately hoping Treasury lifts — and lifts dramatically — its forecasting accuracy if this year’s spending and revenue are to bear even the faintest resemblance to what he will announce tonight. So poor has Treasury’s budget forecasting been in recent years that the figures unveiled tomorrow are unlikely to have much to do with what actually happens in terms of revenue and economic activity in 2015-16.

It’s well known that Treasury seriously underestimated revenue in the years prior to the financial crisis, and since then has consistently overestimated revenue. This shows how each year’s budgets forecast for revenue (including both tax receipts and non-tax revenue) compare to what ended up in government coffers.

There weren’t too many complaints when Treasury serially underestimated revenue, of course, except from some critics who suggested Treasury was doing it to undermine momentum for income tax cuts. But since 20010, Treasury has made life miserable for Wayne Swan and, now, Joe Hockey: each year, the budget has on average forecast over $16 billion more revenue than was eventually booked by the Commonwealth. And just between last year’s budget and the Mid-Year Economic and Fiscal Outlook, Treasury had to write down this year’s revenue as well, by $6 billion.

How much more revenue is written down for this year, and whether it reaches the $16 billion level of the Labor years, will be revealed tonight, along with how much expectations about the coming year’s revenue have been revised down.

The numbers also show how treasurers reacted to Treasury’s errors. As $40 billion-plus a year more than estimated rolled into the Howard government’s coffers, Peter Costello, who these days likes to parade as some sort of paragon of fiscal virtue, reacted by spending most of it; throughout the noughties, Costello’s spending each year reliably blew out by nearly as much as revenue forecasts. Wayne Swan had to take the opposite approach: when revenue came in just under $6 billion below forecast in 2009-10, he cut spending by a little; in 2010-11, when he got around $20 billion less than Treasury predicted, he cut spending by nearly $9 billion. But the following years, he gave up trying to offset falling revenue out of concern for the state of the economy. The financial year 2012-13 came in nearly $18 billion below forecasts, but the final spending result was a $3 billion increase in spending on what the budget for that year had forecast.

Now Joe Hockey — who never accepted that Swan had a revenue problem — has done the same thing. The financial year 2013-14 came in nearly $16 billion below forecasts, but Hockey increased spending by $15 billion. And in 2014-15, without any excuse of being able to blame Labor, MYEFO showed Hockey spending $4 billion more while revenue would be $6.3 billion less.

But go beyond the revenue and spending numbers and you start to see why Treasury’s revenue forecast went awry. Comparing each budget’s forecasts and the final result for the same year in the budget papers, Treasury repeatedly got its terms of trade forecasts wrong: in the decade to 2011-12, Treasury badly underestimated the improvement in Australia’s terms of trade. In 2003-04, for instance, it forecast 1.75% improvement, only for terms of trade to lift 7%, Its forecast for 2006-07 was for flat terms of trade, but they improved by 6.7%. When it tried to anticipate improvement, it got a little closer: it picked 12.25% in 2005-06, which saw 10.9% improvement — which was at least in the ballpark. But its 16% forecast in 2008-09 was way off — they only improved 9.6%.

Treasury was even more off on domestic business investment. During the noughties it persistently underestimated business investment, including 2005-06 when it predicted 6% growth and got 16.2% growth. This is one of the key reasons why it so badly underestimated tax revenue, which continued to pile up year after year before the financial crisis. As late as 2011-12, it forecast a 16% pickup, which became 20%.

But Treasury’s tendency to undershoot on investment has generally become a tendency the other way. In 2012-13, terms of trade were forecast to fall 5.75%, but fell nearly 10%. They were overstated by 3 percentage points in 2013-14. And already this year the forecast fall in terms of trade has doubled from -6.75% to -13.5%. Similarly, business investment growth was only half its 12.5% forecast in 2012-13. The expected 4.5% growth last year became a 5% fall. Even Treasury’s household consumption forecasts, while much closer to the mark — it’s usually between 2.5% and 3.5% — have been overstated in recent years. And GDP has undergone the same pattern: consistent underestimation prior to the financial crisis, persistent overestimation in recent years.

Economic forecasting of course is difficult enough for officials doing their best to anticipate economic trends yet to come — especially in an area like terms of trade that are heavily influenced by foreign economies, global prices and currency movements. That’s before you overlay political judgment. Treasurers are normally presented with a  range of forecasts by Treasury and select from within that range, but there’s nothing to stop a government selecting its own numbers or imposing a consistent preference on Treasury, or deliberately picking underestimates or overestimates to suit the government’s political agenda.

Plainly Peter Costello was in no hurry to correct Treasury’s apparent inability to to stop underestimating revenue, given it made him look good. It was Joe Hockey’s boast with his first MYEFO in December 2013 that the Coalition’s natural prudence in fiscal management would mean an end to the persistent overstatement of revenue by Treasury that occurred during the Labor years. But despite Hockey’s apparent insistence that only the most cautious numbers be used, this hasn’t proven to be the case. December’s MYEFO resulted in $30 billion in revenue write-downs over the next four years on Hockey’s first budget. Tonight may bright  further write-downs on top of those — or might reveal that MYEFO was deliberately understated to serve Hockey’s purposes in producing a better-than-expected deficit tonight.

No one has a functional economic crystal ball, of course, and if they did they are unlikely to be working in either Treasury or politics. The lesson is more that what we are told tonight may end up bearing very little correspondence with what the economic story of 2015-16 really is.